someone recently asked me how i rate movies, so i thought i'd write a little about what ratings mean to me, and make a reference guide. i've had letterboxd since 2017 but for years i resisted rating movies. i felt like i just couldn't commmit, because i knew that once i started rating i'd have to compare every movie to every other movie i'd ever seen, to make sure my ratings system was self-consistent. eventually though, i realised from my own browsing that searching by rating was the best way to find like-minded reviews, and i wanted my reviews to be discoverable too. thus the journey began.
i didn't go into the project of ratings with anything in mind: my system is entirely emergent. my only guiding principle is that it has to be self-consistent, which is not actually as stressful as i thought it'd be. these days i rank my year-in-movies at the end of the year, and i love the tinkering of working out where different movies lie in relation to one another. sometimes this leads to re-rating, as at the end of 2023, when i moved morocco (von sternberg, 1930) from 4.5 to 5 stars, because i realised it had made me crazier than any of my 4.5 star films.
one thing about me is that i'm easily impressed. i love going into movies with no knowledge of what they're about and no expectations (unless i've heard there's body horror, in which case i give it a miss), and i seldom have an actually bad time. as such, it has emerged that my bell curve of ratings peaks at 4 stars. but, at the moment, 3.5 stars is sitting under average for the curve, which i find interesting. the way i interpret this datum is that a movie very rarely makes me go "eh, it was okay." still, i think a lot about the fact that 3.5 stars is a 7/10, which makes it suddenly seem very positive, and this makes me doubt myself.
for this reason i have tried to completely decouple my /5 star system from the /10 or /100 rating systems. 3.5 stars is the border for me, the point where i start feeling "not entirely positive" about something. (not that i won't nitpick at 4 and above too.) so, the more accurate statement would not be that my ratings skew positive, but that i have a rating system that does not reflect traditional ideas of "good" or "bad": it reflects the fact that i just really fucking like movies. it's worth noting that letterboxd has a "like" button, which i have never used, as it adds nothing to my ratings system: the degree to which i like something is directly proportional to how i rate it.
there are so many factors that go into whether or not i like a movie that it's easier to start by outlining what doesn't have any impact on my rating: the big one is production value. under my system, all production tiers are equal. a B-movie could blow me away and a slick blockbuster could leave me cold. actually a lot of slick blockbusters leave me cold haha. i am also not someone who gets super picky about shots angles and shot continuity, unlike some of my friends who have the truly envious skill of an eye for shots. i know and respect a great shot when i see one, but i find it pretty easy to gloss over objectively bad ones. maybe one day i'll have watched enough movies that i understand shots enough for it to matter more to me — i hope so!
a brief word, also, on costuming and set design. i really vastly prefer films that embody naturalism as camus defined it: not how close something is to reality, but how natural the fictional reality is in relation to the characters who live it. to these ends i also tend to prefer verisimilitude, but i don't usually let a lack of it take away from my enjoyment of a film. the issue with costume is that it's far more important to verisimilitude in period pieces, but i can't have two ratings systems depending on whether a movie is period or not. i only let costume (including hair and makeup) impact my rating if it really is egregiously bad; and, usually, this comes alongside with a multitude of other failings, so it becomes a moot point. set design, conversely, is something that only tends to influence me positively. mediocre set design is neutral to me, but good set design is a huge plus, and i can think of multiple examples where it's significantly raised my rating.
and now, the ratings. for each tier, i've tried to give a number of examples that are relatively well-known and show the range of what i'm willing to try. you can find other stuff i love on the vision board page, and all of my letterboxd entries sorted by rating here.
definition: a five-star film is both the easiest and hardest category to define. because: i'll know it when i see it. the way i explain this to friends is that a five-star film "lights up my brain." five stars is firing on all cylinders. i have to like the story, the acting, the pacing, the lighting, the color grading, and of course the cinematography. but it also has to have that extra secret sauce: some kind of insight you aren't getting anywhere else. some transcendent beauty (in the cinematography or otherwise), some unexpected moment of shock, an engrossing performance like no other. often all three, but i particularly like to be shocked, because genuine shock is so hard to achieve. it also has to speak to my well-articulated tastes. a film can be objectively great but it's just not "for" me, and that usually makes it a 4.5. some genres that reliably populate my the five star range are crime, thriller, ensemble cast dramas, dark comedies/satires, and things made in the 70s.
examples: all about eve (mankiewicz, 1950); mikey and nicky (may, 1976); a room with a view (ivory, 1986); sexy beast (glazer, 2000); phantom thread (anderson, 2017)
definition: typically, i give 4.5 stars to films that are generally accepted to be very good, and i agree that they are very good. they don't typically have the secret sauce, but i enjoy them thoroughly. this sounds passionless — it's not. often 4.5-star films do have a dash of the sauce, but there's something else that takes away from it: for example, it could be anything from the few flashes of untempered sentimentality of the otherwise splendid la chimera (rohrwacher, 2023) to the unavoidable misogyny in blow-up (antonioni, 1966), which i genuinely still think is one of the best pieces of art ever made. the key point is that a 4.5-star movie isn't necessarily "almost perfect" — it could be perfect, but it would still be missing the dash of imperfection that makes it what i'm looking for.
examples: the misfits (huston, 1961); suspiria (argento, 1977); pretty woman (marshall, 1990); the royal tenenbaums (anderson, 2001); everything everywhere all at once (sheinert & kawn, 2022)
definition: four stars is my most populous category, and with good reason. these are films i like but do not love, which, when it comes down to it, describes how i feel about most films. typically with a 4-star film i see and appreciate the vision but i have a few issues with the execution; usually pacing or acting, or the film's choice of what to emphasise. or sometimes it's just that i recognise it's a good work of art but it didn't sing to me. often these are films that i find some degree of inscrutable, and while i enjoyed the experience of watching, i left not really knowing what to take away from it, or, if i could identify a takeaway or two, that those themes and meanings didn't satisfy me.
examples: gilda (vidor, 1946); alphaville (godard, 1965); the age of innocence (scorsese, 1993); dogtooth (lanthimos, 2009); anatomy of a fall (triet, 2023)
definition: this is the point at which i'd stop saying "oh i really liked that film." in this way, you could read 3.5 stars as the nebulous borderlands between "films i like" and "films i don't like." it's of course never so absolute, but the rule of thumb is that 3.5 is still a solidly enjoyable experience, with two possible versions of enjoyment: so-bad-it's-good — the west german indie gem fucking city (lambert, 1982) is a fantastic example of this — or well-made-but-dull. a perfect example of the latter category is the zone of interest (glazer, 2023) — glazer is one of my favourite directors and his visuals routinely blow my mind, but this one just plodded along and never quite delivered. these kinds of contradictions are inherent to the strange rating of 3.5. this rating can also mean "i see the vision but i don't respect it," à la saltburn (fennell, 2023), the ultimate in glossy trash carried by incredible acting. 3.5: swings and roundabouts.
examples: the women (cukor, 1933); the apartment (wilder, 1960); bridget jones's diary (maguire, 2001); interstellar (nolan, 2014); the zone of interest (glazer, 2023)
definition: three stars is where i start being actively unhappy with a film. often this is because it is fun but totally lacking in thematic depth and/or technical mastery; often because it's just boring; and sometimes because it's fine but it's really and truly just not my thing. there are also three-star films that i think are actually good, but are ruined by one or two narrative/artistic-choice moments that make you say "how the fuck did this get okayed?" ... for example, the bullshit ending of the technically brilliant close encounters of the third kind (spielberg, 1977), or just about every second scene in the eminently watchable annette (carax, 2021), the only modern musical with good music (there i said it). usually i never feel like i wasted my time with a three-star film, but i do start to feel like some of the people involved in making it were having their time wasted.
examples: what ever happened to baby jane (aldrich, 1962); raiders of the lost arc (spielberg, 1981); my own private idaho (van sant, 1991); inland empire (lynch, 2006); pride and prejudice (wright, 2005)
definition: now we enter the "oh that was a waste of two hours" zone. this is the point at (and below) which i finish a movie and think i could have gone my whole life without seeing this and i would have been happy. i mean, not really — i usually like gaining the cultural knowledge even if i don't enjoy something — but 2.5 is where things start becoming a slog. typically for reasons of either pacing or plot stupidity, these are movies that i find hard to get through. usually i also have limited respect for their cinemacraft, though not always: the graduate (nichols, 1967) is an example of a movie where i thought the cinematography was extraordinary but found the acting wooden and the narrative both uncompelling and unpleasant. however, even in a 2.5-star movie, there may be something to like. a stopped clock is right twice a day.
examples: suddenly, last summer (mankiewicz, 1959); the graduate (nichols, 1967); the last of the mohicans (mann, 1992); the killing of a sacred deer (lanthimos, 2017); licorice pizza (anderson, 2021)
definition: a 2.5-star film is typically not so bad it's funny; two stars is where we start getting into "comically terrible" territory. the thing that makes a film comically bad is that it fails to fully ask the viewer to suspend disbelief. of course, all films from fantasy to realism can ask us to suspend disbelief for any reason, but this must be backed up by the craft. the film must buy into the story it's telling. at two stars, we start to encounter films that don't seem all that invested in their own narrative/emotional reality, and this is something i just cannot abide. at two stars i start to feel that the film has become disconnected from whatever point it's trying to make to the point of incoherence. you might catch me laughing, but i assure you, it's out of astonishment. at two stars, craft becomes irrelevant. we are in the wilderness now.
examples: a matter of life and death (powell & pressburger, 1964); bad lieutenant (ferrara, 1992); bram stoker's dracula (coppola, 1992); ocean's eleven (soderbergh, 2001); elvis (luhrmann, 2022)
definition: by this point, i haven't rated enough films this low to really find strong patterns. but what i will say for almost all of my 1.5-star movies is that i was bored as hell the entire time. not even moved to anger or incredulity: not moved at all. one exception to this is the big lebowski (coen, 1998) where i was moved to the deepest frustration at how badly the film fumbled three good premises and one steve buscemi. but i think on the whole, the thing that kicks in at 1.5 stars is bone-deep tedium. typically at 1.5 and below i also find the craft mediocre, though river of no return (preminger, 1954) shows us some truly genuinely gorgeous shots of the canadian rockies before shitting all over them with a boring script and outrageous misogyny, even for the 50s.
examples: double indemnity (wilder, 1944); river of no return (preminger, 1954); the big lebowski (coen, 1998); jackie (larrain, 2016), darkest hour (wright, 2017)
definition: why did i watch this. why did any of this happen. why did this movie get made. one star films are so bad it's almost impressive that they managed to make it without at least accidentally stumbling into something compelling. how did nobody stop at any point and say "this is so unappealing in every single way, maybe we should just pretend this never happened." typically, this ill will is localised in bad acting and bad screenwriting, which can mar even the most inoffensive cinemacraft. and above all else, these films tend to have the stupidest plots on god's green earth.
examples: ladies' man (mendes, 1931); the devils (russell, 1971); armageddon (bay, 1998); beauty and the beast (condon, 2017); moonfall (emmerich, 2022)
definition: there is something worse yet than the movie you wish had never been made, and that is the movie that you're glad was made because everyone needs something to hate. there needs to be art so worthy and deserving of scorn, so perfectly suited to being ranted and raged at, panned and picked to pieces by buzzards. without bad art, we would never be able to make comparisons; we would never be able to identify the good. genuinely: thank you. i am so fucking angry. but now i have something to talk about.
examples: crocodile dundee (faiman, 1986); florence foster jenkins (frears, 2016); kingsman: the golden circle (vaughn, 2017); mary poppins returns (marshall, 2018); cats (hooper, 2019)